A Collection of Strange Beliefs, Amusing Deceptions, and Dangerous Delusions

From Abracadabra to Zombies | View All

Soal-Goldney experiment

Surprisingly, it is not only believers who are reluctant to imagine fraud, but virtually all skeptics as well ... thus we find skeptics searching for every other conceivable sort of explanation. While the one explanation that is simplest and most in accord with everyday experience is dismissed as inconceivable. -- George R. Price

...unless the accounts of the experiments are completely inaccurate, persons in addition to Soal must have assisted in the deception at some of the sittings. -- C.E.M. Hansel

The Soal-Goldney experiment (1941-1943) was intended to be a replication of the precognitive abilities of Basil Shackleton but turned out to be a replication of dishonesty by a scientist.

S. G. Soal (1889-1975) was a British mathematician. By 1939 he had tested over 160 subjects for ESP in more than 128,000 card-guessing trials and had found no evidence of telepathy or clairvoyance. Whately Carrington advised Soal to recheck his data for displacement (Alcock 1981: 154). Someone with psychic powers might be one or two cards ahead or behind in the guessing because of "temporal distraction." The data mining, conducted by Carrington, Robert Thouless, and Mrs. K. M. Goldney, produced two cases that were statistically significant (Proceedings of the Society for Psychical Research 46, 1940-1941). The two stars were Mrs. Gloria Stewart, who did significantly better than chance at naming the card ahead and the card behind the target card in the test series, even though that was not her task nor her intention (Hansel 1989: 100). The other star, Shackleton, had also been unimpressive in tests done in 1938. The "high displacement scores that Soal alleges he found three years later were the first indication of any striking effects in Shackleton's score sheets" (Hansel 1989: 101).

In January 1941, Soal and Goldney of the Society for Psychical research began testing Shackleton's psychic abilities. Their protocols were elaborately designed and gave the impression of providing very tight safeguards against cheating either by the subjects or the experimenters. Observers from the academic world were present at most of the trials. However, no experts in conjuring participated in the tests. The experiment consisted of 40 sittings of 50 guesses each. The last test was conducted in April 1943.

In one sitting, Shackleton's success at guessing one card ahead was so great that Soal calculated the odds against chance to be greater than 1035 to 1. In another, the odds against chance were calculated to be 1011 to 1. There were several other sessions in which Shackleton performed at phenomenal levels when measured against chance expectation. When a procedure was introduced that sped up the process of testing, Shackleton performed significantly above chance levels at guessing the card two ahead of the target. It looked as if parapsychology had solid scientific proof of psychic ability.

Skeptics were immediately distrustful of the results. Some, like C.E.M. Hansel and George Price (1955), accused Soal of fraud even though they had no direct evidence to support the accusation. They proposed various ways that Soal could have cheated. There was no question that the data were off the charts, "dazzle shots" (to quote Gary Schwartz). But it would take some time before anyone could provide strong evidence of cheating.

In 1954, Soal's Modern Experiments in Telepathy was published. There he reported on his experiments and seemed to brag that he hadn’t bothered with “ultra-rigorous precautions on fraud” by the experimenters because, he said, if the experimenters “are not to be trusted, then there is no point whatever in their doing experiments” (quoted in Christopher 1970: 30). In a classic example of misdirection, when the honesty of Soal's work was questioned he vigorously defended the integrity of Mrs. Stewart, Mr. Shackleton, and the academics who observed his experiments (Christopher 1970: 30; Ciba Foundation Symposium on Extrasensory Perception published in 1956 by Little Brown and Company; and Proceedings of the Society for Psychical Research 53, 1960).

For some time the Soal-Goldney experiment was hailed by many as an example "of the strength of evidence for the reality of ESP" (Thouless). The philosopher C. D. Broad wrote: "Dr. Soal's results are outstanding. The precautions taken to prevent deliberate fraud or the unwitting conveyance of information by normal means are described in great detail, and seem to be absolutely water-tight" (quoted in Hansel 1989: 106). G. Evelyn Hutchinson, a biology professor at Yale University wrote: "Soal's work was conducted with every precaution that it was possible to devise" (quoted in Hansel 1989: 106). J. B. Rhine compared Soal's work favorably to his own sloppily conducted experiments.

Today, it is generally recognized that Soal altered and faked the data, probably unbeknownst to Mrs. Goldney (Alcock 1981: 141). His fraud did significant damage to parapsychology (Alcock 1981: 140-141). Yet, there are still those who hail his work as having produced “the most cogent of all the evidence brought forth to demonstrate the theory of psychic abilities in mankind” (an admirer on the Internet). But those familiar with the history of parapsychology recognize Soal's deception. Dean Radin, for example, in his elaborate and selective defense of the scientific evidence for psi (The Conscious Universe) does not even mention Soal.

the procedure

Most of the sittings were held at Shackleton's photography studio. Two of the sittings were held at the office of the Society for Psychical Research. Most of the tests used the following protocols. In the sender's room, the experimenter sat across from the sender (called "the agent") at a table with a screen that had a small square hole cut in the center. The cards used bore pictures of animals on one side: an elephant, a giraffe, a lion, a pelican, and a zebra. Shackelton - henceforth referred to as S - was to guess what card the agent in another room was looking at. The rooms were adjoining with a door between them that was kept ajar but which did not allow S to see the agent or the experimenter working with the agent. The experimenter would call out to S when it was time to make a guess. Before each trial, the cards, one of each animal, were hand-shuffled by the experimenter or an observer and placed face down and adjacent to each other in a box with an opening toward the agent. Any other observers in the room could not see the cards. The experimenter would signal to the agent which card to turn over by holding up a number from 1 to 5 so that the agent could see it through the hole in the screen. (The card to the far left of the agent was 1 and the others were numbered sequentially from left to right.)

S recorded his guesses on a scoring sheet, numbered 1 through 50. "He was watched throughout by a second experimenter" who checked to make sure he wrote his guesses on the right lines of the sheet (Hansel 1989: 102).

After each run, the cards were turned over by the experimenter working with S and witnessed by S and any observers present. The position of the cards was recorded so the guesses could be compared to the targets.

The numbers for each of the 50 trials in a session were not chosen randomly for each trial, however. They were given to the experimenter by Soal on prepared scoring sheets that Soal "kept under lock and key until brought to the sitting in a suitcase that was never out of his sight" (Hansel 1989: 102). Soal claimed the sheets contained series of numbers chosen randomly, but no protocol was established to observe how Soal actually produced the series of numbered score sheets. This would allow Soal to introduce an excessive number of ones that could later be changed to fours or fives. In fact, one of Soal's agents, Mrs. Albert, testified that she had seen Soal changing a 1 to a 4 or a 5 several times (Medhurst 1971; Hansel 1989: 111; Alcock 1981: 140; Scott and Haskell 1973).

In 1960 Soal and Goldney reported that one "agent" in some of the sittings, "Mrs. G. A.", had told Goldney that she had seen Soal "altering the figures" during the sittings, changing "1's" into "4's" and "5's". This public report by Soal and Goldney came only after Scott had interviewed "Mrs. G. A." and had threatened to publish an account of the matter himself. (Alcock 1981: 140)

Medhurst charged that "the lists of random numbers had not been prepared in the manner stated by Soal and Goldney in their report" (Hansel 1989: 111).

Another nail was driven in by Scott and Haskell in 1974. Even though they could not examine the original score sheets—Soal claimed they'd been lost in 1945—handmade copies of the originals were available. Scott and Haskell examined the score sheets for the sessions when Mrs. Albert served as agent and Soal as the experimenter. They predicted that if the lists were prepared with an excess of 1s that could be changed later they should find, among other things, (a) an excess of 4s and 5s among the targets; (b) a deficiency in the number of cases with a guess of 4 or 5 against a target of 1; and (c) an excess of hits on 4s and 5s  together with an above chance score on 1s (since a 1 would only be changed to a 4 or 5 if it was a miss). All the predicted effects were found in the sitting where Mrs. Albert had made her allegation against Soal and for two other sittings with other agents involved.

The fraud didn't involve a whole lot of work on Soal's part. For example, in session 16 with Mrs. Albert, "it was only necessary to change a 1 into a 4 or 5 about three times in each column [of seven score sheets] to bring about the observed scores" (Hansel 1989: 113). Of special interest is the fact that S would record his guess with a letter representing the animal card (L for lion, for example), not a number, yet "there was no tendency for high scores to arise on particular letters. It was only present in relation to the numbers...." (Hansel 1989: 113).

The final nail was driven in by Betty Marwick in 1978. She confirmed that Soal had not used the method of random selection of numbers as he had claimed.

First, she found two sequences of nineteen digits from two different sittings that matched. A further case was then found, involving the same two sittings where a run of twenty-four digits was involved. In other cases, two series matched when one of them was taken in reverse order. Eventually, following a computer search, it was found that there were frequent cases of matchings of this nature, many of which were not exact, but in which one of the series had extra interpolated digits. These interpolated digits almost invariably secured hits. (Hansel 1989: 114)

Marwick showed that there had been manipulation of the score sheets and that therefore "all the experiments reported by Soal had thereby been discredited." His meticulous experimental design was an illusion (Hansel 1989: 115).

See also clairvoyance, displacement, ESP, experimenter effect, medium, parapsychology, precognition, psi assumption, psychic, remote viewing, telepathy and A Short History of Psi Research by Robert Todd Carroll.

further reading

books and articles

Alcock, James E. (1981). Parapsychology: Science or Magic? Pergamon Press.

Blackmore, Susan. (2001.) "What Can the Paranormal Teach Us About Consciousness?" Skeptical Inquirer, March/April.

Christopher, Milbourne (1970). ESP, Seers & Psychics (Thomas Y. Crowell Co.

Hansel, C.E.M. (1966). ESP: A Scientific Evaluation. Charles Scribner's Sons.

Hansel, C.E.M. (1989). The Search for Psychic Power: ESP and Parapsychology Revisited. Prometheus Books.

Markwick, B. (1978). "The Soal-Goldney experiments with Basil Shackleton: new evidence of data manipulation." Proceedings of the Society for Psychical Research, 56, 250-280.

Medhurst, R. G. (1971). "The Origin of the Prepared Random Numbers Used in the Shackleton Experiments." Journal of the Society for Psychical Research. 46: 44-45.

Scott, C., & Haskell, P. (1973). "Normal" explanation of the Soal-Goldney experiments in extrasensory perception." Nature, 245(5419), 52-54.

Pratt, J. G. (1974). "Fresh light on the Scott and Haskell case against Soal." Proceedings of the Society for Psychical Research, 56(209), 97-111.

Price, G. R. (1955). "Science and the Supernatural." Science, 26 April.

Scott, C., & Haskell, P. (1974). "Fresh light on the Shackleton experiments?" Proceedings of the Society for Psychical Research, 56(209), 43-72.

Scott, C., & Haskell, P. (1975). "Fraud in the Shackleton experiments: A reply to critics." Journal of the Society for Psychical Research, 48(766), 220-226.

Soal, S. G. (1960). "A reply to Mr. Hansel." Proceedings of the Society for Psychical Research, xiiii, 43-82.


Evaluation of Program on Anomalous Mental Phenomena by Ray Hyman

The Experimental Study of Survival and The Problem of the Reality of ESP by Robert H. Thouless

Last updated 12/09/10

© Copyright 1994-2016 Robert T. Carroll * This page was designed by Cristian Popa.